deepDiffusion

A Deep Neural Network-based diffusion equation solver using TensorFlow.

Contributors

• Sayan Adhikari, UiO, Norway. @sayanadhikari
• Rupak Mukherjee, PPPL, USA.

Installation

Prerequisites

1. python3 or higher
2. git
3. conda

Procedure

Using Anaconda/Miniconda

First make a clone of the master branch using the following command

git clone https://github.com/Neural-Plasma/deepDiffusion.git

Then enter inside the deepDiffusion directory

cd deepDiffusion

Now create a conda environment using the given environment.yml file

conda env create -f environment.yml

Activate the conda environment

conda activate deepDiffusion

Usage

Run the code using following command

To train the model

python deepDiff --train

To test the model

python deepDiff --test

Parameter Setup

Edit the input.ini and run the code again. The basic structure of input.ini is provided below,

;
; @file		input.ini
; @brief	deepDiffusion inputfile.
; @author	Sayan Adhikari <sayan.adhikari@fys.uio.no>
;         Rupak Mukherjee <rupakm@princeton.edu>


[grid]
# box size, mm
w = 10.
h = 10.
# intervals in x-, y- directions, mm
dx = 0.1
dy = 0.1

[par]
# Thermal diffusivity of steel, mm2.s-1
D = 1.

[time]
# Number of timesteps
nsteps = 101
# time step to get data from dnn
dnn_start = 50

[dnn]
# number of neurons
nn = 100
epochs = 500
patience = 50
batch_size=32
nlayer = 6

[figures]
plot_fig = True
use_latex = True
add_labels = True

[diagnostics]
dumpData = True

Contributing

We welcome contributions to this project.

1. Fork it.
2. Create your feature branch (git checkout -b my-new-feature).
3. Commit your changes (git commit -am 'Add some feature').
4. Push to the branch (git push origin my-new-feature).
5. Create new Pull Request.

License

Released under the MIT license.